Title Insect-Flower Relationship in the Temperate Deciduous
Total Page:16
File Type:pdf, Size:1020Kb
Insect-flower Relationship in the Temperate Deciduous Forest Title of Kibune, Kyoto : An Overview of the Flowering Phenology and the Seasonal Pattern of Insect Visits INOUE, Tamiji; KATO, Makoto; KAKUTANI, Takehiko; Author(s) SUKA, Takeshi; ITINO, Takao Contributions from the Biological Laboratory, Kyoto Citation University (1990), 27(4): 377-464 Issue Date 1990-08-20 URL http://hdl.handle.net/2433/156100 Right Type Departmental Bulletin Paper Textversion publisher Kyoto University Contr. biol, Lab. Kyoto Univ,, Vol. 27, pp. 377-463 Issued 20 August 1990 Insect-flower Relationship in the Temperate Deciduous 'Forest of Kibune, Kyoto: An Overview of the Flowering Phenology and the Seasonal Pattern of Insect Visits' Tamiji INouE, Makoto KATo, Takehiko KAKuTANi, Takeshi SuKA and Takao IT[No ABSTRACT In 1984 -1987, insect visitors to fiowers werebimonthly or weekly surveyed on a total of 115 plant species or 49 families in the temperate deciduous forest of Kibune, Kyoto, Japan. Flowering was observed from early April to early November, The number of plant species that concurrently bloomed was nine to 17 from May to September. Themonthly total number of flowering plant species peaked twice in May (34 spp.) and September (33 spp,). From April to August, floweringperiods werestaggered arnong congeneric woody species, e.g., Lindera, Rubus, Hydrangea and Deutzia. A total of 4603 individuals of 889 species in 12 orders of Insecta and 2 orders of Arachnoidea were collected. The most abundant order was Hymeno- ptera (46 O/o of the total number of individuals), and it was followed by Diptera (30 O/o) and Coleoptera (140/o). The number of species was highest in Diptera (37 O/o), followed by Hymenoptera (32 O/o) and Coleoptera (17 O/o). The numbers of both species and individuals ofall the insect groups peaked in June and September. There were six farnilies, 14 genera and 73 species in Apoidea. Abundant families were Halictidae (36 O/e of individuals), and followed by Bombinae (24 O/o), Xylocopinae (18 O/o), Andrenidae (8 O/o), Hylaeinae (5 O/o) and Nomadinae (5 O/o), Andrenidae and Nomadinae appeared only before August, whereas Halictidae, Xylocopinae, Bombinae and Apinae were active throughout the fiowering season. Cluster analysis separated 49 plant families into five groups: 13 families were mainly visited by Hymenoptera, 17 by Diptera, five by Coleoptera and two by Hemiptera. The other 12 families were visited by various insect groups. Flowers visited by hymenopterans were further separated into plant families mainly visited by Bombinae, Xylocopinae, Halictidae, Apinae, Xylocopinae, Megachilidae, Vespoidea, Ichneumonoidea and others. Flower preference was compared among insect orders and among families, Themostpreferred plant family was Saxifragaceae in Lepidoptera, Coleoptera and Hymenoptera,Labiatae in Hemiptera and Umbelliferae in Diptera. Flower preference greatly varied among families in Apoidea. The most preferred plant family was Umbelliferae in Colletidae, Saxifragaceae in Halictidae and Andrenidae, Compositae in Megachilidae, Balsaminaceae in Bombinae, and Rosaceae in Apinae. KEY WORDS flowering phenology/anthophilous fauna/floral host/Kyoto/ temperate deciduous forest introduction • We studied flowering phenology and community structure of flower visiting insects in 1984-87 at three locations in Kyoto Prefecture, Japan. This paper is a preliminary report of the study at Nbune. The other two locations were Ashu (Kato et al., 1990) i Contribution to the ecological and biolconomical studies of the pollinator commumty in Kyoto M. 378 INouE et al. and Kyoto City (Kakutani et al., 1990). "Ihe study area in Kibune is a mosaic of natural temperate deciduous forests and planted evergreen coniferous forests. Kibune is intermediate between the primary cool-temperate beech forests in Ashu and the disturbed urban area in the evergreen oak forest zone in Kyoto city. This paper describes the relationship between flowers and flower visiting insects mainly at family or subfamily level. rlhe community structure of flower visiting bees has been studied at seven localities with different latitudes in Japan: Sapporo (43e 02'N, Sakagami and Fukuda, 1973), Hakodate (410 46'N, Matsumura and Munakata, 1969), Nikko (36e 45'N, Nakamura and Matsumura, 1985), Gifu (saO 39'N, Yamauchi et al,, 1974), Sasayama (350 04'N, Miyamoto, 1ee2), Wakayama (340 03'N, Matsuura et al., 1978) and Kochi (33e 33'N, Ikudome, 1978). We will compare the bee community in Kibune with these data. Detailed analysis of insect-flower relationship at species level will be done in other papers. Study Site Kibune (35e 10'N, 1350 52'E) is located at about 10 km north to the urban area of Kyoto City (Fig. 1), Its elevations range from 300 to 740 m abovesea level. This area contains primary forests that are well conserved due to the religious regulation (Fig. 2). There are also planted forests of ChryPtomeria 1'aPonica. The climax stage on mountain ridges in this area is the evergreen coniferous forest that is predominated by Abies firma and Tsuga sieboldii or the deciduous oak forest predominated by Quercus crisPula and Acer spp. Gravely slopes along streams are covered with the secondary walnut (luglans mandshurica) forests that are accompanied by Acer carPinifolium, darPinus laxiflora, C, tschonoskii and EuPtelea Polyandra. Floors of the walnut forest are covered with shrubs such as CePhalotaxus harringtonia, Deuzia maximowicziana, Hydrangea hirta, H. macroPhylla and Lonicera gractliPes and a shrubby herb LeucoscePtrum stelliPilum, Open areas along streams are occupied by shrubs such as Deutzia crenata, Rubus spp., PVeigela hortensis and riparian herbs, Persicaria thunbergii, Reynoutria 1'aPonica, ImPatiens spp., Angelica PolymorPha, A. Pubescens and Ligularia fischerii. Kibune has been famous in rich flora for botanists. From Mt. Kurama, near the study site, a total of 1422 native species of vascular plants (Pteridophyllum 140 spp., Gymnospermum 15 spp. and Angiospermum 1267 spp.) have been recorded (Hiroe, 1973). Climate of this area is warm ternperate. Only incomplete meteological data are available at Hanase pass which is located at just 1 km north of the study site (35e 9'N, 1350 47'E, 710 m above sea level, Fig. 3). There was much rainfall in June and July but very little in August in 1984 to 1987. From December to March, north wind brought frequent snowfalls and the deptrh of snow was usually 10 to 60 cm in this period. Although we have no detailed data on air temperature of our study site, Insect-flower RelationshiP in Kibune 379 * Pass 6bb ? x t/ '`'t' 8 "" lj q' 8 Fig. 1. The study area in Kibune, Kyoto and its location in the Kinki District, Japan (inset). The streamside along A and B is in Asoga stream and that along B and C is in the Kibune river. Flower visitors were collected on flowring plants along C - B - A. An asterisk shows the location of the meteological observatory at the Hanase Pass, it was in general 5 OC lower than that at the campus of Kyoto University (Elevation= 60 m, Kakutani et al., 1990) and 1 OC higher than that at Ashu (Elevation = 400m, Kato et al., 1990). Methods We chose, for samplings, fine days during the flowering season from April to early November in four years from 1984. Census intervals were about two weeks in 1984 - 1986, and one week in 1987. ln each census day, one or two of us walked on the fixed route (3 km) along the Asoga stream and the Kibune river (Fig. 1). We made quantitative samplings of insect visitors on flowers (detailed methods in Kato et al. 1990). Samples included real nectar/pollen users (pollinators and rubbers), ambush 380 INouE et al, ffEevss y,gvaee Fig. 2. Landscape of the study area along the Asoga stream. (a) a view of the study area, (b) a natural d eciduous forest right and a planted ChryPt- omeria forest left near site B , (c) riparian vegetation along a stream near site A. Insect-flower RelationshiP in Kibune 381 600 predators for flower visitors, feedersofflow- AE er petals and plant 1eaves and insects that E arejust resting near flowers. Thesecategories were distinguished in Results and Discussion. v 400 =U From 1984 to 1986, sampling started -c between 9:OO and 10:OO h. It took up to oro six hours to finish the census, depending -År 200 on the amount of flowering plants. In 1987, Åí samplings were made twice both in morning 8 and afternoon (start between 12:oo-13:oo h) 2 o in the same day. Sampling dates are as follows: 30 1984: May 22, June 28, July 10, 23, 20 mÅr August 11, 23, September 7, 15, October o 10 1, 12 a 1985: April 4, 21, May 15, June6, 20, July o 8, August 11, September 7, 8, October 8 M J J A S O 1986: April 30, May 22, June 4, l2, July Month 24, August 5, September 26, October 8 Fig. 3. Changes in monthly rainfall (mean 1987: April 23, May 9, 18, 30, June 18, in 1984-87, bar = s.d.) and the weather conditions•(the mean 27, July 9, 22, 31, AugrLst 1, 8, 26, 27, number of days per month in September 10, 19, 27, October, 3, 10, 20 1984-87), These meteologicaldata were recorded at Hanase Pass meteological observatory (710 rn The flowering period of each plant species above sea level, Fig. 1, Kyoto Prefectural Meteological reports, was estimated as the mean of the flowering 1984-1987). period directly observed and the interval from the latest date at which flowering had not yet started to the earliest date at which flowering had finished. All arthropod specimens were labeled with complete census data (date, locality and flower species). They were classified and identified at species level by specialists who are shown in Acknowledgments. Statistical analyses were done by the SAS pack- age in the Data Processing Center, Kyoto University.